改进果蝇算法优化BP神经网络预测汽油辛烷值  

Optimizing BP neural network to predict gasoline octane number by improved fruit fly algorithm

作　　者：韦修喜 陶道 黄华娟[1] WEI Xiuxi;TAO Dao;HUANG Huajuan(School of Artificial Intelligence,Guangxi Minzu University,Nanning 530006,Guangxi,China;College of Electronic Information,Guangxi Minzu University,Nanning 530006,Guangxi,China)

机构地区：[1]广西民族大学人工智能学院,广西南宁530006 [2]广西民族大学电子信息学院,广西南宁530006

出　　处：《山东大学学报（工学版）》2023年第5期20-28,36,共10页Journal of Shandong University（Engineering Science）

基　　金：国家自然科学基金资助项目(62266007,61662005);广西自然科学基金资助项目(2021GXNSFAA220068,2018GXNSFAA294068);广西研究生教育创新计划项目(JGY2022104)。

摘　　要：针对BP神经网络存在预测精度不佳、预测结果不稳定的问题,提出改进果蝇算法优化BP神经网络(back propagation neural network)预测模型。将混沌映射、判别因子与变步长机制引入果蝇优化算法(fruit fly optimization algorithm,FOA)中,得到改进后的自适应混沌果蝇优化算法(fruit fly optimization algorithm with chaos and discriminant factors,CDFOA),并利用测试函数对算法进行性能验证。利用CDFOA优化BP神经网络的初始权值与阈值,构建基于CDFOA优化BP神经网络对于汽油辛烷值的预测模型CDFOA-BP。将采集到的60组汽油数据输入预测模型进行测试分析。预测结果表明,与FOA-BP模型、PSO-BP模型、SSA-BP模型和BP神经网络模型相比,CDFOA-BP模型在预测精度与预测稳定性上均优于其他4种模型,验证该模型的有效性与可行性。Aiming at the problems of poor prediction accuracy and unstable prediction results of back propagation(BP)neural net-work,a prediction model based on improved fruit fly algorithm and optimized BP neural network was proposed.Chaos mapping,discriminant factor and variable step size mechanism were introduced into the fruit fly optimization algorithm(FOA)to obtain an improved fruit fly optimization algorithm with chaos and discriminant factors(CDFOA),and the performance of the algorithm was verified by function test.The initial weights and thresholds of BP neural network were optimized by CDFOA,and the prediction model CDFOA-BP for gasoline octane number based on CDFOA optimized BP neural network was constructed.60 groups of collect-ed gasoline data were input into the prediction model for test and analysis.The prediction results showed that compared with FOA-BP model,PSO-BP model,SSA-BP model and BP neural network model,CDFOA-BP model was superior to the other four models in prediction accuracy and prediction stability,which verified the effectiveness and feasibility of the model.

关 键 词：果蝇优化算法 混沌映射 判别因子 函数测试 BP神经网络 辛烷值 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]